EDITORIAL STANDARD

Editorial & Scenario Policy

This site is designed to be frightening to read at times. That makes source discipline more important, not less.

The rule: dramatic framing is allowed. Fabricated evidence, fake certainty and invented emergencies are not.

Three labels govern our content

DOCUMENTED

Documented fact

Supported by a primary source, official guidance, credible research, a directly attributable report or a clearly identified dataset.

SCENARIO

Scenario analysis

A plausible chain of events built from known capabilities, dependencies or historical failure patterns. A scenario is not a prediction.

SPECULATIVE

Speculative extreme

A deliberately severe thought experiment used to test resilience. It must never be written as though it is already happening or certain to happen.

How scenario articles should be written

State the triggering assumption clearly.
Separate what is known from what is inferred.
Identify major dependencies and points where the scenario could fail to unfold.
Do not invent casualty figures, dates, government warnings or probabilities.
Do not use fake countdowns, fabricated alerts or manipulated imagery that could be mistaken for documentation of a real event.
End with practical resilience actions that remain useful outside the scenario.

Fear, urgency and headlines

Headlines may be provocative because the publication is intentionally built around catastrophic risk. The body must earn that headline. We will not claim that extinction, collapse or an AI takeover is certain when the evidence does not support certainty.

Health, emergency and cyber material

Medical, food-safety, water-safety and emergency guidance carries higher stakes. We prefer official public-health and emergency-management sources, avoid personalised diagnosis or treatment, and tell readers when official local instructions override general guidance. Cyber content is defensive: protection, recovery and resilience rather than instructions for unauthorised access or harm.

Commercial independence

Advertising, affiliate commission, email conversion and product sales are never acceptable reasons to alter a factual conclusion. We do not place ads or offers so they masquerade as emergency instructions, download buttons or navigation.

Updating fast-moving AI claims

AI capabilities change quickly. Material claims about model behaviour, cyber incidents, regulation and frontier-system evaluations should carry dates and be rechecked when a page is materially updated. Old scenarios may remain useful, but their evidence layer must not be presented as current when it is not.

Last reviewed: 21 August 2026